STUDY OBJECTIVE--The aim was to investigate risk factors for cancer of the buccal and labial mucosa in Kerala, southern India. DESIGN--The investigation was a case-control study. SETTING--Regional Cancer Centre, Trivandrum, Kerala, and local teaching hospitals. PARTICIPANTS--Cases were all those registered with oral cancers at the Regional Cancer Centre during 1983 and 1984 (n = 414). Controls (n = 895) were selected from admissions to the cancer centre who were found to have non-malignant conditions, or from patients attending outpatients in teaching hospitals of Trivandrum medical college with non-malignant conditions. MEASUREMENTS AND MAIN RESULTS--The risk in males of the following habits was investigated: pan (betel)-tobacco chewing, bidi and cigarette smoking, drinking alcohol, and taking snuff. Only pan-tobacco chewing was investigated in females as very few indulged in other habits. Among males predisposing effects were found for pan-tobacco chewing (p less than 0.001), bidi smoking (p less than 0.001), drinking alcohol (p less than 0.001), and taking snuff (p less than 0.01). As in males, pan-tobacco chewing also had a predisposing effect in females (p less than 0.001). Duration of use was a better predictor of risk than either daily frequency of use or total lifetime exposure, both for pan-tobacco chewing (especially if the habit started before age 21 years) and bidi smoking. However, there were also very high risks associated with the current occasional use of both factors. Pan-tobacco chewing was the most important risk factor, with relative risk of 13.24 with 31-40 years' use, and 37.75 with greater than 40 years' use among males. Corresponding relative risks in females were 21.30 and 54.93. No effect of cigarette smoking was observed (relative risk 0.64, p greater than 0.1). CONCLUSIONS--A substantial majority of cases of buccal and labial cancers are attributable to chewing pan-tobacco. This has obvious implications for instituting preventive measures.